five

open-llm-leaderboard-old/details_h2oai__h2ogpt-oasst1-512-12b

收藏
Hugging Face2023-10-16 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_h2oai__h2ogpt-oasst1-512-12b
下载链接
链接失效反馈
官方服务:
资源简介:
该数据集是在Open LLM Leaderboard上对模型h2oai/h2ogpt-oasst1-512-12b进行评估时自动创建的。数据集由64个配置组成,每个配置对应一个被评估的任务。数据集由2次运行生成,每次运行都作为每个配置中的一个特定分割存储。train分割始终指向最新的结果。此外,还有一个results配置存储了所有运行的聚合结果,用于在Open LLM Leaderboard上计算和显示聚合指标。

该数据集是在Open LLM Leaderboard上对模型h2oai/h2ogpt-oasst1-512-12b进行评估时自动创建的。数据集由64个配置组成,每个配置对应一个被评估的任务。数据集由2次运行生成,每次运行都作为每个配置中的一个特定分割存储。train分割始终指向最新的结果。此外,还有一个results配置存储了所有运行的聚合结果,用于在Open LLM Leaderboard上计算和显示聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总

数据集概述

数据集简介

该数据集是在评估模型 h2oai/h2ogpt-oasst1-512-12bOpen LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

  • 该数据集包含 64 个配置,每个配置对应一个评估任务。
  • 数据集从 2 次运行中创建,每次运行可以在每个配置中找到一个特定的分割,分割名称使用运行的时间戳。
  • "train" 分割始终指向最新的结果。
  • 一个额外的配置 "results" 存储所有运行的聚合结果,用于计算和显示 Open LLM Leaderboard 上的聚合指标。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_h2oai__h2ogpt-oasst1-512-12b", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-16T18:51:34.628441 运行的最新结果: python { "all": { "em": 0.0010486577181208054, "em_stderr": 0.000331458146521924, "f1": 0.049975880872483294, "f1_stderr": 0.0012253223818797603, "acc": 0.339436730966812, "acc_stderr": 0.00841008969581668 }, "harness|drop|3": { "em": 0.0010486577181208054, "em_stderr": 0.000331458146521924, "f1": 0.049975880872483294, "f1_stderr": 0.0012253223818797603 }, "harness|gsm8k|5": { "acc": 0.016679302501895376, "acc_stderr": 0.0035275958887224733 }, "harness|winogrande|5": { "acc": 0.6621941594317285, "acc_stderr": 0.013292583502910887 } }

配置详情

  • harness_arc_challenge_25

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|arc:challenge|25_2023-07-19T18:11:10.994515.parquet
  • harness_drop_3

    • 分割:2023_10_16T18_51_34.628441
    • 路径:**/details_harness|drop|3_2023-10-16T18-51-34.628441.parquet
    • 分割:latest
    • 路径:**/details_harness|drop|3_2023-10-16T18-51-34.628441.parquet
  • harness_gsm8k_5

    • 分割:2023_10_16T18_51_34.628441
    • 路径:**/details_harness|gsm8k|5_2023-10-16T18-51-34.628441.parquet
    • 分割:latest
    • 路径:**/details_harness|gsm8k|5_2023-10-16T18-51-34.628441.parquet
  • harness_hellaswag_10

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hellaswag|10_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hellaswag|10_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_abstract_algebra_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-abstract_algebra|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_anatomy_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-anatomy|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_astronomy_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-astronomy|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_business_ethics_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-business_ethics|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_clinical_knowledge_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-clinical_knowledge|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_biology_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_biology|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_chemistry_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_chemistry|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_computer_science_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_computer_science|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_mathematics_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_mathematics|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_medicine_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_medicine|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_college_physics_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-college_physics|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_computer_security_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-computer_security|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_conceptual_physics_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-conceptual_physics|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_econometrics_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:11:10.994515.parquet
    • 分割:latest
    • 路径:**/details_harness|hendrycksTest-econometrics|5_2023-07-19T18:11:10.994515.parquet
  • harness_hendrycksTest_electrical_engineering_5

    • 分割:2023_07_19T18_11_10.994515
    • 路径:`**/details_harness|hendrycksTest-electrical_engineering|5_2023-07-19T18:11:10.994
二维码
社区交流群
二维码
科研交流群
商业服务